View source for The Framework Convention on AI and human rights, democracy and the rule of law – TOPIC 03 Sub 02 2024

19 June 2024 | 11:30 - 12:15 EEST | Auditorium | [[image:Icons_live_20px.png | Video recording | link=https://youtu.be/7-W1bJbbJkM]] | [[image:Icon_transcript_20px.png | Transcript | link=The Framework Convention on AI and human rights, democracy and the rule of law – TOPIC 03 Sub 02 2024#Transcript]] <br />
[[Consolidated_programme_2024#topic3_sub2_24|'''Consolidated programme 2024 overview''']]<br /><br />
Proposals: #31 #36 #39 #70 #72 #82 (see [https://www.eurodig.org/wp-content/uploads/2024/01/EuroDIG-2024_List-of-proposals-20240116_for_wiki.xlsx list of proposals])<br /><br />
== <span class="dateline">Get involved!</span> == 
You are invited to become a member of the Session Org Team by simply subscribing to the [https://list.eurodig.org/mailman/listinfo/topic3_sub2_2024 '''mailing list''']. By doing so, you agree that your name and affiliation will be published at the relevant session wiki page. Please reply to the email send to you to confirm your subscription.

Kindly note that it may take a while until the Org Team is formed and starts working.

To follow the current discussion on this topic, see the [[{{TALKPAGENAME}} | discussion]] tab on the upper left side of this page.

== Session teaser ==
On 17 May 2024, the Council of Europe adopted the Framework Convention on Artificial Intelligence and human rights, democracy and the rule of law, the first international treaty in this field, which will be opened for signature in Vilnius in September. This session offers a discussion on this first-of-its-kind instrument with those directly involved in its elaboration.

== Session description == 
The Council of Europe Framework Convention on Artificial Intelligence and human rights, democracy and the rule of law is the first-ever international legally binding treaty aimed at ensuring the respect of human rights, the rule of law and democracy legal standards in the use of artificial intelligence systems.

Its adoption by Committee of Ministers of the Council of Europe (Ministers of Foreign Affairs) on 17 May 2024 marked the culmination of several years of intense work within the Committee on Artificial Intelligence (CAI), involving the 46 Council of Europe member states, the European Union and 11  non-member states (Argentina, Australia, Canada, Costa Rica, the Holy See, Israel, Japan, Mexico, Peru, the United States of America, and Uruguay), as well as representatives of the private sector, civil society and academia, who participated as observers.

The aim of the session is to bring together different stakeholders having taken an active role in the elaboration and negotiation of the Framework Convention, and to offer the audience an insight into the making of this groundbreaking instrument.

== Format == 
The session is divided into 2 parts: 

-	a panel discussion between the key participants moderated by the Chair of the CAI, Ambassador Thomas Schneider

-	interaction with the audience, in the form of a Q&A

== Further reading == 

Information on the Framework Convention can be found on its [https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence dedicated webpage], and in this newly published [https://rm.coe.int/ai-convention-brochure/1680afaeba brochure] (available also in [https://rm.coe.int/brochure-convention-ia/1680afaebb French])

== People == 
'''Please provide name and institution for all people you list here.'''

'''Programme Committee member(s)''' 
*Desara Dushi
*Jörn Erbguth
*Meri Baghdasaryan,
*Minda Moreira
The Programme Committee supports the programme planning process throughout the year and works closely with the Secretariat. Members of the committee give advice on the topics, cluster the proposals and assist session organisers in their work. They also ensure that session principles are followed and monitor the complete programme to avoid repetition.

'''Focal Points''' 
*Louise Riondel, Council of Europe
*Swiss Federal Office of Communications (OFCOM)
Focal Points take over the responsibility and lead of the session organisation. They work in close cooperation with the respective Programme Committee member(s) and the EuroDIG Secretariat and are kindly requested to follow [https://www.eurodig.org/get-involved/planning-process/#tab-organising-a-session EuroDIG’s session principles] 

'''Organising Team (Org Team)''' ''List Org Team members here as they sign up.''
*Gianluca Diana
*Giacomo Mazzone 
*Louise Riondel
The Org Team is a group of people shaping the session. Org Teams are open and every interested individual can become a member by subscribing to the mailing list.

'''Key Participants'''

*Mr David FAIRCHILD - First Secretary, Permanent Mission of Canada to the Office of the United Nations, Head of Delegation of Canada to the CAI

*Mrs Laura GALINDO - AI and Privacy Policy Manager, META 

*Mr Mario HERNÁNDEZ RAMOS - Constitutional Law Professor, Complutense University of Madrid, Head of Delegation of Spain to the CAI

*Mr David MARTI - Founder, AI Programme Lead, Pour Demain (online)

'''Moderator'''

*Mr Thomas SCHNEIDER - Ambassador, Chair of the CAI, Vice-Director, Swiss Federal Office of Communication (OFCOM), Swiss Federal Department of the Environment, Transport, Energy and Communications (DETEC)

'''Remote Moderator'''

Trained remote moderators will be assigned on the spot by the EuroDIG secretariat to each session. 

'''Reporter'''

Reporters will be assigned by the EuroDIG secretariat in cooperation with the [https://www.giplatform.org/ Geneva Internet Platform]. The Reporter takes notes during the session and formulates 3 (max. 5) bullet points at the end of each session that: 
*are summarised on a slide and  presented to the audience at the end of each session 
*relate to the particular session and to European Internet governance policy
*are forward looking and propose goals and activities that can be initiated after EuroDIG (recommendations)
*are in (rough) consensus with the audience

== Current discussion, conference calls, schedules and minutes ==
See the [[{{TALKPAGENAME}} | discussion]] tab on the upper left side of this page. Please use this page to publish:
*dates for virtual meetings or coordination calls
*short summary of calls or email exchange
Please be as open and transparent as possible in order to allow others to get involved and contact you. Use the wiki not only as the place to publish results but also to summarize the discussion process.

== Messages ==  
The CoE Framework Convention on AI and human rights, democracy and the rule of law is an important step towards a global approach to AI regulation. The CoE Framework and the EU AI Act compliment each other. Further steps should follow, by taking into account the need to address the growing issues of AI from a global, rather than a regional perspective.

== Video record ==
https://youtu.be/7-W1bJbbJkM

== Transcript ==
'''''Disclaimer:''' This is not an official record of the session. The DiploAI system automatically generates these resources from the audiovisual recording. Resources are presented in their original format, as provided by the AI (e.g. including any spelling mistakes). The accuracy of these resources cannot be guaranteed.''

[https://dig.watch/event/eurodig-2024/main-topic-3-the-framework-convention-on-ai-and-human-rights-democracy-and-the-rule-of-law '''Transcripts and more session details were provided by the Geneva Internet Platform''']


Thomas Schneider:
Unfortunately, the gender balance is not really optimal. We were to have a lady from META, but she felt sick, so she canceled yesterday, unfortunately. So, yeah. Well, welcome to this second session this morning on topic three, on AI. And this is mainly, but not only, focusing on one of the key instruments of these thousands of instruments that we’ll see, we’re seeing and will continue to see in the future. It’s, of course, the Convention of the Council of Europe. We’ve had our, the Secretary General Maria Pechenovic-Buric talk about it, and me as well in the previous round. So, it has been adopted by the Committee of Ministers on 17th of May. And as we’ve heard, it will be open for signature in September. We hope that it’s quite an intense timeline. In the CHI, we’ve been used to intense timelines, but that’s life. Things get faster. So, we’ll hope that, yeah, at least some countries will be ready to sign it in September. And we do have two people here that were participants in the negotiations from representing governments, from different ministries, and, of course, from different continents, as Canada is not yet part of Europe formally. So yeah, we should let you into the Eurovision Song Contest thing either. Last night proved that there’s talent, so that’s definitely something to think about. But again, of course, it is important to note that the And that’s what the Secretary-General has said. There were big discussions during the negotiations on whether we should have a European Convention that would be much more based on and oriented and using the European instruments that we have. Of course, mainly the European Convention on Human Rights and the Court and everything that is based on this versus trying to go global, trying to develop an instrument that is attractive to also non-European countries. And of course, there are pros and cons and there were hard fights and people were criticising me in the media, for instance, to be too inviting to non-European countries. But I was convinced, and so was the Secretariat, that it is important to try and develop something that may have a global reach, given that most of the AI systems are developed in countries that are not part of the European Union or not even part of Europe. And we do have shared values, but different ways in implementing them. So I’m very happy again to have two people from two different continents here. So let me turn to our first panellist, my dear friend Mario from Spain. You’re a law professor, so he knows actually the legal part much better than I do, and he was very useful in this during the negotiations in Madrid. And he was the head of the delegation of Spain in the negotiations, and a member of the Bureau as well. Very constructive support. Thanks for this again. What are your takeaways? What did you learn? To what extent were your expectations fulfilled with this convention? How do you see it in the bigger landscape? I know that you also were in Latin America a number of times talking to people from that continent, because you speak Italian so well. Of course, that helps. What are your takeaways? What did you learn and what is coming in the future? 

Mario Hernandez Ramos:
Good morning. Can you hear me? First of all, Thomas, thank you very much for the invitation. Hello, dear Thomas. I would like to say my first words just to thank you once again, Thomas, because you were in charge of the guy and you made a very good job because, believe me, the guy was a very, very challenging task. And I did believe that we couldn’t make it because the time was very tight, the complex things were very challenging, and you did a very terrific job. And thanks to the secretariat also, with Christian, Badim and Luis, because they did a very good job. So they were the best people. And in the end, you succeeded. You drove us to the end in this very happy ending. The Framework Convention, I think, is very good news, if you see also along with the European Union Regulation on AI Act, because these two instruments are complementary. And this is a very important idea, and you were very keen on having other countries out of Europe because we have different perspectives, different legal backgrounds, but the AI is so common, the challenges are very common. So you had from the very beginning a very clear idea that only European countries could make general standards. So we need to be there also Canada for sure, United States, Japan, Israel, other countries, Latin American countries. So the first idea I would like to highlight is that these two European instruments are complementary. And this is good news, because without the European Union regulation on AI, there was no Framework Convention of the Council of Europe. Both instruments, we need to take them together. And my second idea I would like to highlight is regarding democracy, because I’m a constitutional lawyer, so I was very worried about the democracy challenges regarding artificial intelligence, because the European Union regulation was very focused on – well, not very focused, but the main concern was fundamental rights, but not about democracy or rule of law. And democracy is one of the main topics regarding the Council of Europe, and we were working very hard from the very beginning about democracy, and it was not very easy to try to regulate through a legally binding instrument a very complex institution as democracy. And by complex, I mean not difficult, one by an institution composed by many institutions – principles, fundamental rights, processes – and as a matter of fact, we have to be clear that artificial intelligence is playing a political role already in these three elements that we could allocate in this guide regarding democracy, like the confirmation of public opinion, the electoral processes, and the representative institutions. And artificial intelligence is playing an active role in all of those elements. So trying to develop a legally binding instrument regarding these three institutions was so challenging, so challenging. And from a point of view, we could allocate one fundamental principle to face the two main and most problematic characteristics of artificial intelligence regarding democracy, which are autonomous capacity of taking decisions and self‑learning. Which is the individual autonomy principle. So since we are human beings, we need to be considered as individuals. So we need to work that every human being is treated like a person, like an individual human being. And not like part of a group or not to be associated as a part of a profile. This is the main task of this from my point of view and from a constitutional lawyer point of view of this Framework Convention. So individual autonomy is a fundamental principle which is there in Article 7, also in Article 1, and in the spirit of the whole Framework Convention. Thank you. 

Thomas Schneider:
Thank you. And I’m very happy that you raised, because most of the people are talking about human rights when it comes to AI, also when it comes to the convention, also the convention itself is much more specific on human rights than on democracy. But if you remember the Menti result, actually disinformation, whatever you call it, information integrity was the highest, let’s say, concern. And we also had lots of discussions and also lots of expectations on how to regulate democracy or protect democracy from issues of AI. So I hope that that will come up in the discussion, because it’s actually a very challenging issue. So let me now turn to the Mr. David Fairchild. He is the first Secretary of the Permanent Mission of Canada to the Office of the United Nations in Geneva, and was the head of delegation of Canada to the CAI. Very outspoken, so keeping reminding us about how many hours we have left. from the very beginning to the end to negotiate that treaty. And yeah, how was it for you? Why is Canada participating at the Council of Europe where you are only so-called observer and why engage in a cumbersome process with European countries? 

David Fairchild:
Good morning. Hi, good morning, everybody. You can all hear me well. I do do moonlighting as a backup singer for Thomas at international events, but primarily here to talk about the AI treaty. So I guess let me start by thanking, first of all, Thomas, the Council of Europe for the opportunity. Christian, you’re in the, I can see you, but thank you. It was a long two-year journey that’s sort of coming to an end, I think, but nevertheless, to your question, I mean, I think Canada was a part of, so prior to the launch of the negotiations, there was another process prior to that, which was more of a working group that spent a couple of years studying this question before they launched the negotiations, which Canada was a part of. And I think the reason we were a part of this process is fundamentally, I think we recognized that there was a need, either normative, regulatory, or legislative, to address the growing issues around AI. And I think one of the challenges, and you’ve heard this over the last couple of days, is we talk about AI, we go from existential risks of killer robots and Arnold Schwarzenegger breaking through the door at any moment to the very positive benefits that AI are going to deliver. Sitting in Geneva, you have a lot of countries who hear both ends of the spectrum and frankly have nothing to hang anything on. And so there’s a real gap. I mean, emerging technology generally is a massive gap. And I think there’s recognition that there’s a gap, and states, regional bodies, at an international level, we are simply attempting to move to try to address some of that. So why we wanted to be there was fundamentally, this was going to be the first international legal instrument. on artificial intelligence. And I think that probably is the key takeaway for everybody here. This is the first international legally binding treaty that is going to regulate signatories on AI. And I think that’s a very fundamental statement that I think is very important that Canada want to be a part of. Secondly, and I think this goes to the terms of reference of the whole process, is that they wanted to not offer a completed negotiated text and treaty to the world. They wanted the world to be part of the process. And I think that was both an opportunity and challenge for the Council of Europe, and I think also for the European Union, without saying too much on that, simply because this question of complementarity versus alignment came up throughout the entire negotiation. And I say this with the most humblest of statements, that as an observer, our job was to ensure that the convention itself remained as applicable as possible to the greatest number of states. And that presented some very key challenges in the negotiation by the simple fact that between Europe and other parts of the world, there are some 99% symbiosis, but some areas where there are some clear differences. And our job was to challenge the process to ensure that that space was left open for other countries, whether it be Canada or someone else down the road, and to ensure that the people of the world can find themselves in the treaty. And to understand the treaty is a very slim document. It’s what we call a framework treaty. And so ultimately, the treaty basically says you need to have a cat to catch mice. They don’t care what color the cat is. They don’t care if it has a long tail or no tail, or long hair or short hair. It has to catch mice. And that’s the fundamental of the framework treaty. is that cat. For other countries, they were at very different stages of development of that framework at the national level. as I quote my UK colleague where normally when you’re doing international treaties you are leveling up national legislation To a higher level at the international level in this space most countries didn’t have something to start with and so we were setting the Bar, and that was a very difficult challenge in terms of the negotiation because there are some concepts some ideas You know whether it be you know again, it’s late. That’s a legally binding text, so I think that was one of the the major challenges, but Ultimately why Canada wanted to be there was to send a strong signal to the world You know there are people who are going to respect human rights democracy in the rule of law And there are those who challenge that notion and so for us being part of that process and being a signatory hopeful signatory to this Treaty is something. I think we want to stand by 

Thomas Schneider:
Thank you very much David we were supposed again to have a Laura Galindo from Meta Joining us, but unfortunately she felt sick so she couldn’t be here, and then there’s another Person that was supposed to come which is David Marti From a Swiss from an NGO that was participating also in the negotiations But he hasn’t which is unusual probably he’s stuck somewhere in the traffic jam In a tunnel in a Swiss train that is not working because otherwise He would probably be here the good thing But this is that we have more time to actually have a discussion among among the the people here so I would say let’s jump right into the discussion and Raise your hands online or on site and then yeah floor is open 

Audience:
I Have some thoughts of I mean, I think the question is, do you think that this approach, this convention, is really balancing innovation and regulation to, I mean, learn from the motto of this year’s EuroDIG? No, I’m saying this because we, also thinking at the previous panel, we had this very long discussion on, let’s never use AI for migrants because that could, I mean, possibly, indeed, that could lead to, I mean, bias and not a proper treatment of their rights. And I, while I understand this and I support this, I was thinking that in my past experience in Italy, I saw migrants that had to wait sort of like 12 months in detention for a judge to be available to judge their application for a refugee status. And so, I mean, maybe technology can also do good things. So maybe, yes, we cannot use technology to speed up things. And the result could be, I think, more infringing on the person’s human rights than if we had used the technology with all the risk of biases. So the point, in my view, should be not rather, I mean, let’s not use these technologies, but how can we use these technologies in a way that preserves the human rights and addresses the inevitable problems and failures that technology can have in some cases. So I was wondering whether there was this kind of approach, because it’s very easy to start to think at a high level, you know, how do we protect everyone’s rights and, but the risk is then to never do anything. 

Thomas Schneider:
Thank you. I think let’s take a few interventions and then so that we are not always the ones here speaking. So there’s somebody online, I understand. Yep. You can go. 

Audience:
Thank you. Can you hear me? 

Thomas Schneider:
Yes, we can hear you, but not. Thank you. Go ahead. Introduce yourself and go ahead. Thank you. 

Audience:
Thank you. Advocate Isanuba from South Africa, University of South Africa, lecturing in IP, AI, and cyber law. With some stint in international law, international human rights. Have been in the military in South Africa. I’m raising this because I want to place a context of my question. Prof, thank you, from Spain, in sketching the mischief behind the coalition. Then my question, having been following on the developments, when the EU-AI Act was actually put in action, the question I asked you was regarding a clause that is there in the data privacy space, which deals with data adequacy, which is critical when one is dealing with the issues as you know. I’m going to ask the same thing on the convention, whether I ask the relationship between the convention on AI and the convention that deals with data privacy. And there’s a reason I’m asking that. As you might know, how the multilateral legal instruments flow, and Africa then would then develop an instrument, according to which then the domestic laws are developed. Having listened to Canada, I would want you to respond in terms of firstly, what is the relationship between this convention and the convention that deals with data privacy, because all of them are dealing with the same thing. The second thing is, how do we then see the rollout of the multilateral legal instrument? And what would be the requirements in terms of its enforceability? And I’m asking this, being aware of the GDPR, and how GDPR has began to actually move beyond the convention in terms of its efficacy. Thank you. even in Africa, because then you are contracting with the EU. So those are the two questions, but I’m quite excited to be participating in this, because I’m also doing a lot of writing and research and supervision in the space of AI and IP. Thank you very much. This is Advocate Zaluba from South Africa. 

Thomas Schneider:
Thank you very much, and this shows also how interconnected we are across the continents. Whatever we do in one part of the world may have consequences in the other. So let’s take one more, and then maybe let… 

Audience:
Johann Erbgut, University of Geneva. I’ve got two short questions. One question concerning Mario. You said this Framework Convention wouldn’t have been possible without the AI Act. That this Framework Convention wouldn’t have been possible without the AI Act. Many countries, including Switzerland, don’t have the AI Act or something similar. And I think, isn’t this framework something that comes before, for many countries, before having a concrete Act? And the other question is concerning the bias. We have often discussed bias as something that is coming with training data, that is automatically produced by self-learning systems. And now we are confronted with positive bias. Positive bias, as we can see in Google Germany, where we see that the answers are directly biased in a positive way, that society likes. But shouldn’t we have a right to be free of this positive bias, to have something that is more neutral and not positively biased? So, what do you think about this intentional bias that is increasingly being introduced into the systems, so the systems… are good, but, of course, then they are biased. Thank you. 

Thomas Schneider:
Okay. Let’s give two colleagues here a chance to reply to Okay. 

Mario Hernandez Ramos:
Thank you very much. So kind, as usual, you are. Starting with the question about that we needed the European Union regulation to be passed before to pass or to agree on the Framework Convention. This Framework Convention contains very general principles, very general principles. Not all the countries are, I mean, maybe non-countries are very happy with the end because we wanted to be a little bit more ambitious regarding fundamental rights and protection. But we have to give up some pretensions because we understand that, for sure, most of the countries, we don’t have, even Spain or normally all the countries, we don’t have national regulation on that. So many governments were against having first a law or a legally binding norm, an international one before the national one. So they wanted to have a detailed legally binding law from the national perspective and after that, the international one. But I mean, that is the scenario we had. And we couldn’t have this Framework Convention because, for instance, we wanted like specific Bill of Rights regarding artificial intelligence, for instance, right? We have already. those Bill of Rights in all the constitutions, national constitutions, and also in the European Union Regulation, now we have it all to be protected. So what I’m trying to say, I’m sorry because I’m a bit nervous, it’s like I am sometimes because the thing is like in order to have some countries that they wanted to have first detailing national regulations the only way to have them on board were to accept very general principles, and these general principles we were very we were agreed on, we have an agreement on that, but some countries didn’t want to have a specific regulation because it was very difficult to sell them back home. So that was the only one thing, the only thing, the only way of doing things, yeah, so and complementary things. Data protection, of course, this is one of the main topics, how to complement these international instruments, and also about data protection. I’m pretty sure that you are almost passing a law, I think, on artificial intelligence? 

David Fairchild:
Yeah, no, thanks Mario. So I guess there’s an interesting counterpoint, so did we need the EU AI Act to have an international convention in a different regional body? No. That would be my short answer, and that was one of the arguments I made quite often in the room, and it didn’t necessarily please the European Commission who was negotiating on behalf of the EU, but when you invite non-EU member states to a Council of Europe treaty process, my considerations are what’s important for Canada, and frankly, a European body of law is not interesting to Canada when I’m signing an international treaty. So there was, this is just a clear example of some areas where there was, I wouldn’t say conflict, but there was certainly some cheeky discourse in the room between friends. And I think that’s part of the, I think part of the interesting part of this whole process. You know, we had United States, Canada, Israel, Japan, Mexico, and then later on a number of South American countries, including Argentina, Colombia, Peru, who joined the negotiations later on. But what you realized throughout this process is that everybody was at different points. And to your answer, and I think the online question is Canada’s not a signatory to 108+. We are currently in the process, two years in, trying to pass an updated piece of legislation to address AI and data and privacy. There are different opinions about whether we will be successful before we go into an election. But this was a reflection of the problem. You know, the EU could dominate, in a way, the conversation substantively, but there were other countries. You know, the United States does not have legislation. Canada is in the process of developing legislation. And so I think what it resulted in was that we had to focus on the things where we could find agreement. So it was perhaps reductive in that sense, but nevertheless, it was creating a common start point that where all of the entities could agree. And on the question that was raised, innovation was actually one of the very important topics that featured heavily through the negotiations, Vittorio, because I think the question always is gonna be is how do you protect rights while protecting economic growth and innovation? And so there was a lot of time spent trying to figure out how do you regulate AI in its development stage? How do you protect R&D from regulations while still respecting the need for ethical and rights-based approach? And so I think we found, hopefully found a landing spot, but often. As a diplomat, I would say negotiations aren’t about being right. Negotiations are about finding a landing spot where everybody can agree. I think there was a very strong recognition that this was merely a first step, but we needed to take it as urgently as possible to try to create an international leveling up that was binding on member states. I think that’s the crucial piece here. Nothing else that you are talking about at the international level, whether it be in the UN or elsewhere, is binding on member states. And so, when you sign a treaty, you are taking on legal obligation. 

Mario Hernandez Ramos:
And regarding innovation, I think there was an endless debate, and there’s a kind of contradiction between innovation and protection of human rights, but I think this is a false contradiction because innovation must be done from the protection of human rights. So if you want to make an innovation, you cannot make it trying to destroy human rights. It’s like if you want to innovate about cars running through people because you want to show how strong is the metal or whatever. So you need to make the innovation taken into account by default the respect of human rights. Because if you want to make an innovation without taking into account human rights, the outcome will be not very respectful, of course, with human rights. So from my point of view and from all point of view, it’s a false dichotomy. It’s not a problem of respecting human rights. It’s a way or a certain way of making innovation. And this is the human-centered AI perspective that we do believe in. 

Thomas Schneider:
I’m very happy that you say this because you’re a law professor. As an economist, I would say we live in capitalist societies. where the stronger one, the more agile one wins and gets rich and the one that is too slow, doesn’t find the right services will disappear. And from a legal point of view, you’re right. But in reality, it is a tension. There is a tension where you have incentives for people to, if they can’t go against human rights, at least, yeah, go to the limits because just as much beyond this, they are not criticized, but it’s not supposed me to be talking. So let’s go back to the audience and react. We have an online hand up. So please let us hear what the person is saying. 

Audience:
So it’s me. Hello, Wolfgang Benedek, University of Graz. First of all, my congratulations to the speedy conclusion of that convention, which I think is really a great achievement. My question would address some criticism found in the reactions to the convention regarding the implementation. Because the advantage of the framework convention binding compared to the code of ethics of UNESCO is its binding nature. But then regarding the implementation, it seems that states are rather free how to do this, whether by binding laws or also in the form of self-regulation of companies. But if that would be the result, then the question is really, what is the value added of a binding convention? Thank you. 

Thomas Schneider:
Let’s take a few more. David. 

David Fairchild:
Very intense negotiations. And as I said just a few minutes ago, sometimes getting a treaty is more important than getting the language right. And sometimes you have to start somewhere. And so this criticism of it was all or nothing, I think what I would respond is that it’s a half loaf is better than no loaf. And so do we recognize that over time this treaty could be improved with additional protocols or other processes? Absolutely. I think what we all recognize is that we needed to get something out the door and that something had to at least be either equally unpleasant or equally pleasant to all potential signatories. 

Mario Hernandez Ramos:
Yeah, completely agree. And also, I mean, about the implementation, the national implementation, I think we need to make a difference between if we are talking about the 27 member states of the European Union, because they don’t need to implement this Framework Convention because they are going to be implemented by the regulation. So there was not going to be any problem with the private sector. The problem is with the countries outside of the European Union. I mean, the problem is how are they going to implement in the private sector? And then it comes, the answer of David, it’s better to have this convention than having nothing. That was what I tried to mean when I got so nervous and I couldn’t explain myself. It’s better to have this general agreement on private sector and to trust that all the countries, Canada, United States, Israel, they want to regulate at least than having nothing. And that is the real achievement of this Framework Convention. But from the 27 European Union member states, we have no problems. That is the relationship. This is the complementary. perspective of these two instruments. From the European Union perspective, and that’s why we couldn’t have this framework convention adopted before having the regulation, we have no problems in this sector, in the private sector. That’s what I wanted to say before, sorry. 

Thomas Schneider:
I think we will discuss that we have no problems, but when you talk to the countries that have to implement not only the AI Act, but the DMA, DSA, Media Freedom Act, and it’s a little bit of a struggle from what I understand. So I wouldn’t say we have no problem, but maybe you’re slightly ahead of others. But that’s another discussion. To pick up what has been said, I think we could maybe focus a little bit still on the democracy part. Also, Wolfgang Benetech’s criticism, we have like two minutes left. Okay, because talking about human rights is easy, but having a binding convention on democracy and on things like what is false information, what is disinformation? Do you really want to have the states being obliged to regulate the media and so on by law when it comes to what is truth or not? I’m simplifying things, but I think that’s a key question that actually we haven’t, we still haven’t solved. The EU has some acts that go in that direction. Many others are discussing, my country would be impossible to have the state tell the media what is right or wrong because we have a different approach. So I think that is a key thing. So maybe one more one-minute intervention and one minute answer from you, and then we have to stop. 

David Fairchild:
Thanks, Thomas, really easy one. What I would simply say is, yes, I just wanted to answer the earlier question about positive bias. I mean, separate to the treaty is another much longer process that’s called the Huderia. So human rights, democracy, rule of law, impact assessment. And so there is an actual quite a long terms being run by the Alan Turing Institute to try to flesh out what would be a generic risk framework that countries could use. And I think some of these questions around how to address bias is being looked at. And so, Canada is part of this process. We’ve come with a case study, in fact, that uses our immigration department, which uses a substantive number of IAI systems. And I’m sure other countries similarly have other case studies. But to your point, so we have a number of pieces of legislation, and all of these touch on aspects of the treaty. But fundamentally, and I think this is just a general takeaway comment, is that we really understood well human rights in the context of this council treaty negotiation. Things kind of fell off fairly quickly when we started to think about what are the rights, what is upholding democracy? Is that parliament? Is that, you know, offices of privacy? And then rule of law. And so, there was sort of this open-ended question that is still sort of, I think, being ruminated on is what does a legally binding treaty mean for upholding democracy and rule of law when there really isn’t a lot of case study or perhaps common understanding? And so, we effectively had to go down to basic principles. 

Mario Hernandez Ramos:
Okay. Ten seconds. Sure. Sure. I mean, maybe we don’t need to regulate media, but we need to bear in mind that artificial intelligence is there. And what we do need is to assure that people can freely form opinions. And for that, we can use legally binding general principles. And judges or other legal operators can apply them case by case. Because regulating is very complex. But knowing in a certain case that that is a misinformation, that is a deepfake, is easier than to regulate from a general perspective. Thank you. 

Thomas Schneider:
So, to summarize, we have now a few things achieved on paper. We have a convention. which is general, but there’s some principles, there’s some obligations in the what, with quite an option in the how, but the what is clear. We do have an AI act and a number of other things that may also influence countries from outside the EU, but these things now exist on paper. Now we need to make all of this work. And that is at least as complex as developing the papers. I think we agree on this, and I also would say we agree on, we are still at the beginning of developing the normative system around how to deal with AI, but at least we have started. I think that’s the main point. Thank you very much. Thank you all for joining. 

[[Category:2024]][[Category:Sessions 2024]][[Category:Sessions]][[Category:Main Topic 3: Artificial Intelligence 2024]]